A method and apparatus for evaluating laser thermal effects

Abstract

The application discloses a laser heat effect evaluation method and device, and belongs to the technical field of biology. The method comprises the following steps: acquiring an original laser speckle blood flow image sequence, analyzing the original laser speckle image by using a time contrast analysis algorithm to acquire a laser speckle contrast image; converting the laser speckle contrast image to acquire a laser speckle blood flow image. The half-height width of the blood vessel in the laser speckle blood flow image is calculated as the width of the blood vessel; the blood flow image after laser irradiation is subtracted from the blood flow image before laser irradiation to acquire a difference image; the blood vessel with blood flow change caused by laser irradiation is acquired according to the difference image, then a region of interest is selected, and the mean value of the blood flow of the blood vessel before and after laser irradiation is acquired, so that the final laser treatment effect is evaluated. The mouse dorsal window 108 is irradiated from the complete skin side of the mouse dorsal window 108, and the blood flow change is monitored in real time by using the laser speckle blood flow imaging technology on the exposed blood vessel side of the opposite side. The influence of the upper skin tissue is considered, the blood flow change is monitored in real time, and the precise evaluation of the laser heat effect on the blood vessel is realized.

Description

TECHNICAL FIELD

[0001] The application provides a laser thermal effect evaluation method and device, and belongs to the technical field of biology. BACKGROUND

[0002] The manufacturing process of the mouse spinal window model is carefully designed to minimize surgical trauma and accelerate animal recovery. Through fine surgical techniques, researchers can accurately create a transparent observation window in the mouse spinal area on the back. This process is not only technically feasible, but also relatively short in operation time, improving the efficiency of the experiment. In addition, the introduction of the transparent window allows the microstructure such as blood vessels to be clearly visible without additional staining or labeling, greatly simplifying the experimental steps and costs.

[0003] In the research of laser therapy, the thermal response of blood vessels is one of the key factors in evaluating the effectiveness and safety of treatment. The mouse spinal window model, through its unique design, allows researchers to directly and real-time observe the changes in blood vessel morphology, blood flow velocity and surrounding tissue thermal effects after laser irradiation. This observation method not only has high resolution, but also can capture instantaneous physiological responses, providing valuable data support for in-depth understanding of the interaction between laser and biological tissue.

[0004] Although the mouse spinal window model has significant advantages in blood vessel observation, in traditional research methods, the treatment laser is often directly irradiated on the blood vessels on one side of the excised skin. This approach to some extent ignores the influence of the upper skin tissue on the absorption, scattering and heat transfer process of laser energy. This neglect may lead to inaccurate evaluation of the thermal effect of deep blood vessels, because the skin, as the first barrier for laser entering the body, its optical and thermal properties will have important influence on the distribution and transmission of laser energy. This leads to the traditional laser thermal effect evaluation method of irradiating laser on the peeled side of the mouse spinal window while monitoring the laser thermal effect, ignoring the influence of the upper skin tissue on absorption, scattering and heat transfer, and cannot accurately evaluate the thermal effect of deep blood vessels. SUMMARY

[0005] In order to overcome the deficiency that the prior art cannot accurately evaluate the thermal effect of deep blood vessels, the application provides a laser thermal effect evaluation method and device, which irradiates the mouse spinal window from the intact skin side of the mouse spinal window and monitors the blood flow changes in real time on the exposed blood vessel side using laser speckle blood flow imaging technology. Both the influence of the upper skin tissue and the real-time monitoring of blood flow changes are considered, and accurate evaluation of the laser thermal effect on blood vessels is realized.

[0006] In order to solve the above technical problems, the technical scheme adopted by the application is:

[0007] In a first aspect, the application provides a laser thermal effect evaluation method, comprising:

[0008] The original laser speckle blood flow image sequence is acquired, and a time contrast analysis algorithm is used to analyze the original laser speckle image to obtain a laser speckle contrast image; the laser speckle contrast image is converted to obtain a laser speckle blood flow index image;

[0009] The half-width of the blood vessels in the laser speckle blood flow index image is calculated as the width of the blood vessels; the blood flow image after laser irradiation is subtracted from the blood flow image before laser irradiation to obtain a difference image;

[0010] According to the difference image, the blood vessels whose blood flow changes due to laser irradiation are obtained, and then a region of interest is selected to obtain the mean value of the blood flow of the blood vessels before and after laser irradiation, so as to evaluate the final laser treatment effect.

[0011] As a further improvement of the application, after the original laser speckle blood flow image sequence is acquired, the following steps are further included:

[0012] The original laser speckle blood flow image is denoised in real time by using deep learning, and is converted into a speckle blood flow image;

[0013] The deep learning is a laser speckle blood flow image denoising generative adversarial denoising method, and the deep neural network of the laser speckle blood flow image denoising generative adversarial denoising method is composed of a denoiser and a discriminator, and the denoising is realized by mutual confrontation between the denoiser and the discriminator;

[0014] The denoiser and the discriminator both adopt Unet architecture; the denoiser is composed of an encoder, a bridge and a decoder; the encoder unit adopts three lightweight residual blocks arranged in sequence, the stride is 1 or 2, is used for feature extraction, and the number of filters is doubled after downsampling operation; wherein, the first layer of downsampling with a stride of 1 is delayed; the extracted features are transmitted to the decoder through the bridge, and are used for estimating the denoised input in the feature domain; the features are converted to the image domain by each decoder layer through a conversion unit, the conversion unit is composed of n residual blocks, followed by a 1x1 convolution layer and a Sigmoid activation layer; the number of filters in the conversion unit is halved for n times, and the filter size of the 1x1 convolution layer is set to C;

[0015] Further comprising a blood vessel structure enhancer, the blood vessel structure enhancer includes a large-core depthwise separable convolution to extract low-level features, and then captures structure information of different scales through a group of parallel depthwise separable convolutions; the structure features are fused through two 1x1 convolution layers, and the blood vessel structure enhancer is attached to the last decoder layer before entering the conversion unit;

[0016] The discriminator network is composed of a normal convolution layer, and the output of the conversion unit in the discriminator is converted into a binary mask through a mapper, and the mapper is composed of a convolution layer and a Sigmoid activation layer.

[0017] As a further improvement of the present application, a contrast signal-to-noise loss function is introduced in the time contrast analysis algorithm to enhance the contrast of blood vessels and tissues by guiding the adversarial distortion learning network.

[0018] As a further improvement of the present application, the laser speckle contrast image is converted to obtain a laser speckle blood flow index image using a sequence formed by the original laser speckle blood flow image.

[0019] In a second aspect, a laser speckle blood flow imaging device is provided, comprising a laser speckle imaging system and a therapeutic laser; the laser speckle imaging system comprises an imaging laser, a beam expander, an imaging optical system, an image acquisition system, a band-pass filter, a neutral density adjustable attenuator, and a mirror.

[0020] The mouse spine window animal model is fixed in the laser speckle imaging system by a fixing device, and the laser generated by the therapeutic laser is reflected by the first mirror and then vertically incident on the side of the mouse spine window complete skin, while the other side of the skin is removed for blood flow monitoring by the laser speckle imaging system.

[0021] The imaging laser selects a near-infrared waveband laser light source, and the laser power is adjusted by changing the angle of the neutral density adjustable attenuator; the adjusted light beam is expanded by a spot beam expander and then irradiates the mouse spine window through the second mirror; the mouse spine window is imaged by the imaging optical system, and the image is acquired by a detector and then transmitted to a data acquisition and analysis system for processing.

[0022] Optionally, an interference band-pass filter with the same center wavelength as the imaging laser is installed in front of the detector to collect dynamic scattering signals; the therapeutic laser irradiates the mouse spine window complete skin side through the mouse spine window.

[0023] In a third aspect, the present application provides a laser heat effect evaluation system, comprising:

[0024] The acquisition module is configured to acquire a sequence of original laser speckle blood flow images, analyze the original laser speckle images using a time contrast analysis algorithm to obtain a laser speckle contrast image, and convert the laser speckle contrast image to obtain a laser speckle blood flow index image.

[0025] The calculation module is configured to calculate the half-height width of the blood vessels in the laser speckle blood flow index image as the width of the blood vessels, and obtain a difference image by subtracting the blood flow image after laser irradiation from the blood flow image before laser irradiation.

[0026] The evaluation module is configured to obtain blood vessels with changed blood flow due to laser irradiation according to the difference image, select a region of interest, obtain the mean blood flow of the blood vessels before and after laser irradiation, and evaluate the final laser treatment effect.

[0027] As a further improvement of the present application, it also includes a noise reduction module:

[0028] The noise reduction module uses deep learning to reduce the noise of the original laser speckle blood flow image in real time and converts it into a speckle blood flow image.

[0029] The deep learning is a laser speckle blood flow image noise reduction generative adversarial denoising method, and the deep neural network of the laser speckle blood flow image noise reduction generative adversarial denoising method is composed of a denoiser and a discriminator, and the denoising is realized by mutual confrontation between the denoiser and the discriminator.

[0030] The denoiser and the discriminator both adopt Unet architecture; the denoiser is composed of an encoder, a bridge and a decoder; the encoder unit adopts three lightweight residual blocks arranged in sequence, the stride is 1 or 2, which is used for feature extraction, and the number of filters is doubled after downsampling operation; wherein the first layer of downsampling with a stride of 1 is delayed; the extracted features are transmitted to the decoder through the bridge for estimating the denoised input in the feature domain; the features are converted to the image domain by each decoder layer through a conversion unit, and the conversion unit is composed of n residual blocks, followed by a 1x1 convolution layer and a Sigmoid activation layer; the number of filters in the conversion unit is halved n times, and the filter size of the 1x1 convolution layer is set to C.

[0031] It also includes a blood vessel structure enhancer, which contains a large-core depthwise separable convolution to extract low-level features, and then captures structure information of different scales through a group of parallel depthwise separable convolutions; the structure features are fused through two 1x1 convolution layers, and the blood vessel structure enhancer is attached to the last decoder layer before entering the conversion unit.

[0032] The discriminator network is composed of ordinary convolution layers, and the output of the conversion unit in the discriminator is converted into a binary mask through a mapper, and the mapper is composed of a convolution layer and a Sigmoid activation layer.

[0033] As a further improvement of the present application, in the acquisition module, a contrast signal-to-noise ratio loss function is introduced in the time contrast analysis algorithm to enhance the contrast of blood vessels and tissues through a guided adversarial distortion learning network.

[0034] As a further improvement of the present application, in the acquisition module, the laser speckle contrast image is converted to obtain a laser speckle blood flow index image using a sequence formed by the original laser speckle blood flow image.

[0035] In a fourth aspect, the present application provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the laser heat effect evaluation method when executing the computer program.

[0036] In a fifth aspect, the present application provides a computer readable storage medium, which stores a computer program, and the computer program, when executed by a processor, implements the laser thermal effect evaluation method.

[0037] In a sixth aspect, the present application provides a computer program product, which comprises computer instructions, and the computer instructions instruct a computer to execute the laser thermal effect evaluation method.

[0038] The present application has the following beneficial effects over the prior art:

[0039] The laser speckle imaging technology adopted by the present application is a non-contact and non-invasive imaging method, which does not cause any harm to the subject and is suitable for wide application in clinical and scientific research. The time contrast analysis algorithm can capture the blood flow changes of the microvessels, so that the method has high sensitivity to the blood flow changes caused by laser irradiation. By calculating the mean value of the blood flow of the blood vessels before and after laser irradiation, the quantitative evaluation of the laser treatment effect can be realized, which provides a scientific basis for clinical decision-making. The laser speckle imaging technology can capture the dynamic changes of the blood flow in real time, so that the method has an advantage in evaluating the immediate effect of the laser treatment process. The method is not only suitable for evaluating the laser treatment effect, but also can be used in the fields of monitoring blood vessel diseases and studying hemodynamics, and has a wide application prospect. The present application proposes to irradiate the mouse dorsal window from the intact skin side of the mouse dorsal window, and to monitor the blood flow changes in real time by using the laser speckle blood flow imaging technology on the exposed blood vessel side of the opposite side. Both the influence of the upper skin tissue and the real-time monitoring of the blood flow changes are considered, so that the precise evaluation of the laser thermal effect on the blood vessels is realized. BRIEF DESCRIPTION OF DRAWINGS

[0040] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following introduces the drawings of the related technical solutions in the embodiments of the present application or the prior art. It should be understood that the drawings in the following introduction are only for the convenience of clearly describing part of the embodiments of the technical solutions of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0041] Figure 1 Fig. 1 is a schematic diagram of a mouse dorsal window 108;

[0042] Figure 2 Fig. 2 is a system diagram of a laser speckle blood flow imaging device of the present application;

[0043] Figure 3 Fig. 3 is a placement physical object diagram of a mouse dorsal window 108 of the present application;

[0044] Figure 4 Fig. 4 is a flow chart of laser speckle blood flow imaging data acquisition and analysis of the present application;

[0045] Figure 5 The blood flow images of the conventional peeling irradiation and the transmission method of the present application are compared.

[0046] In the figure, 100, imaging laser; 101, adjustable attenuator; 102, beam expander; 103, second mirror; 104, therapeutic laser; 105, detector; 106, band-pass filter; 107, imaging optical system; 108, mouse spine window; 109, first mirror; 110, data acquisition and analysis system. DETAILED DESCRIPTION

[0047] The embodiments of the present application are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar notations represent the same or similar elements having the same or similar functions throughout. The embodiments described below by referring to the accompanying drawings are exemplary only, for the purpose of explanation, and are not to be understood as a limitation of the present application. For the step numbers in the following embodiments, they are only set for the convenience of description, and the order between the steps is not limited in any way, and the execution order of each step in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.

[0048] In the description of the present application, the words such as setting, installing, connecting, etc. should be understood in a broad sense unless otherwise explicitly limited, and those skilled in the art can reasonably determine the specific meaning of the above words in the present application in combination with the specific content of the technical solution.

[0049] The first object of the present application is to provide a laser heat effect evaluation method, comprising:

[0050] S1, obtaining a sequence of original laser speckle blood flow images, analyzing the original laser speckle images by using a time contrast analysis algorithm to obtain a laser speckle contrast image; further converting the laser speckle contrast image to obtain a laser speckle blood flow index image.

[0051] S2, calculating the half-width of the blood vessels in the laser speckle blood flow index image as the width of the blood vessels; obtaining a difference image by subtracting the blood flow image after laser irradiation from the blood flow image before laser irradiation;

[0052] S3, obtaining the blood vessels whose blood flow changes due to laser irradiation according to the difference image, then selecting a region of interest, obtaining the mean value of the blood flow of the blood vessels before and after laser irradiation, and thus evaluating the final laser treatment effect.

[0053] The time contrast algorithm,

[0054] ;

[0055] Wherein and respectively, are the standard deviation and mean of the time series of individual pixel points of the original speckle image.

[0056] The laser speckle contrast image is converted to obtain an original laser speckle blood flow index image BFI The formula is as follows:

[0057] ;

[0058] This laser thermal effect evaluation method is mainly based on laser speckle imaging technology and image processing algorithm, which is used to evaluate the influence of laser irradiation on blood flow, and then evaluate the effect of laser treatment. Its principle is:

[0059] Laser speckle blood flow imaging: Laser speckle is a random interference pattern formed by the scattering of micro-particles (such as red blood cells) inside the tissue when laser irradiation reaches the surface of the biological tissue. The change of these speckle patterns over time reflects the dynamic information of blood flow. By capturing these speckle image sequences with a high-speed camera, the dynamic changes of blood flow can be obtained.

[0060] Temporal contrast analysis: By using temporal contrast analysis algorithm to process laser speckle image sequence, the contrast image reflecting the change of blood flow velocity can be extracted. The time contrast is a measure of the speed of speckle pattern change over time, which is proportional to the blood flow velocity. Therefore, the contrast image can directly show the distribution and velocity of blood flow. The laser speckle contrast image is converted to obtain a laser speckle blood flow index image.

[0061] Vessel width measurement: By calculating the half-height width (i.e. the width of the vessel when the brightness reaches half of the peak value) of the blood vessels in the laser speckle blood flow index image, the width of the blood vessels can be estimated. This step helps to more accurately locate and analyze the blood vessel area.

[0062] Difference image generation: Subtract the blood flow image after laser irradiation from the image before irradiation to obtain the difference image. The bright or dark areas in the difference image represent the areas where the blood flow has changed significantly after laser irradiation, which are often the key areas for treatment effect evaluation.

[0063] Laser treatment effect evaluation: Select the region of interest (ROI) in the difference image, calculate and compare the mean value of blood flow in the region before and after laser irradiation. By comparing these values, the influence of laser irradiation on blood flow can be quantitatively evaluated, and the effect of laser treatment can be evaluated.

[0064] Based on the present application, animal experiment evaluation of various medical lasers can be carried out, which is beneficial to precise diagnosis and treatment of vascular and pigmented skin diseases, and has great economic benefits. Based on the patent technology of the present application, the accuracy of laser thermal effect animal experiment can be improved, which is beneficial to obtaining more accurate treatment parameters and improving treatment effect.

[0065] The content of the present application is described in detail below.

[0066] First, the mouse dorsal window 108 is made. When making the mouse dorsal window 108, the back skin is lifted up, and a part with rich blood vessels is selected for fixation. The skin on one side is removed with scissors after being disinfected with alcohol, and is cut into the shape of the circular window 108. Then, it is fixed with a titanium alloy sheet, and is fixed by sewing with a needle, and the finished mouse dorsal window 108 is as shown in Figure 1 .

[0067] As shown in Figure 2 , unlike the conventional direct irradiation method of treatment laser, the present application proposes a new method of transdermally irradiating the mouse dorsal window 108 with treatment laser, observing the subcutaneous blood vessels on the complete skin side, and preferentially using laser speckle blood flow imaging blood flow monitoring means to monitor the blood flow changes in real time during the treatment process. The blood flow monitoring means can be selected from ultrasonic Doppler, laser Doppler, magnetic resonance, etc., and the transdermal irradiation of treatment laser is closer to the real treatment scene, and the imaging laser can effectively observe the blood flow changes.

[0068] Figure 2 A schematic diagram of the laser speckle blood flow imaging device provided by the present application is shown, which mainly consists of a mouse dorsal window 108 animal model and a fixing device, a laser speckle imaging system, and a treatment laser 104. The laser speckle imaging system consists of an imaging laser 100, a beam expander 102, an imaging optical system 107, an image acquisition system, a band-pass filter 106, a neutral density adjustable attenuator 101, and a mirror (103, 109). The imaging laser 100 can be selected from a near-infrared waveband laser light source, and the laser power can be adjusted by changing the angle of the neutral density adjustable attenuator. The adjusted light beam is expanded by the beam expander 102, and is irradiated to the mouse dorsal window 108 through the second mirror 103 at an angle of 60° with the vertical plane. The imaging optical system 107 is used to image the mouse dorsal window 108, and can be selected from a lens and a stereoscopic microscope. The CCD and sCMOS detectors 105 are used for image acquisition, and the processed data is transmitted to the data acquisition and analysis system 110 for processing. In order to prevent the treatment laser from interfering with the imaging laser, the interference band-pass filter 106 with the same central wavelength as the imaging laser 100 is installed in front of the detector 105 to collect the dynamic scattering signal. As shown in Figure 2 , the mouse dorsal window 108 is placed and fixed on the hollow object table, and the side with the removed skin is upward. The laser generated by the treatment laser 104 is reflected by the first mirror 109 installed on the 45-degree mirror holder to receive high-energy laser, and is incident perpendicular to the complete skin side of the mouse dorsal window 108, while the side with the removed skin is monitored for blood flow by the laser speckle imaging system. Figure 1 The middle green light spot indicates the irradiation area of the treatment laser.

[0069] The data processing flow of the present application is shown in Figure 3 First, the original laser speckle blood flow image sequence is obtained by using the self-made laser speckle blood flow imaging device, and the original laser speckle blood flow image is analyzed by using a time contrast analysis algorithm with high spatial resolution to obtain a laser speckle contrast image. The high spatial resolution of this method facilitates the observation of small diameter microvessels.

[0070] Further, in order to improve the time resolution while maintaining the signal-to-noise ratio of the image, the original laser speckle blood flow image is denoised in real time by using deep learning, and is converted into a speckle blood flow image.

[0071] The optional number field anti-distortion learning denoising method comprises a denoiser and a discriminator, and the two achieve high-quality denoising effect by mutual confrontation. The network structure is lightweight designed, and real-time denoising can be achieved.

[0072] In view of the problem that the boundaries of blood vessels and tissues are not obvious, the method introduces a contrast signal-to-noise ratio loss function to guide the anti-distortion learning network to enhance the contrast of blood vessels and tissues. The deep learning denoising method is used to denoise the original laser speckle blood flow image calculated by using only 5 frames, and the quality of the obtained speckle blood flow image exceeds that of the traditional algorithm of 25 frames of original speckle blood flow image, and the time resolution is improved by 5 times, which greatly improves the monitoring ability of the laser-induced transient thermal effect.

[0073] In summary, the laser speckle blood flow imaging device and the deep learning algorithm developed above have high spatiotemporal resolution laser transient thermal effect monitoring ability. Then the obtained high-quality blood flow image sequence is analyzed to evaluate the curative effect of laser treatment.

[0074] The specific steps include: first, the half-width of the blood vessels in the blood flow image is calculated as the width of the blood vessels, then the blood flow image after laser irradiation is subtracted from the blood flow image before laser irradiation to obtain a difference image, and the blood vessels with changed blood flow caused by laser irradiation are obtained, then the region of interest is selected, and the mean value of the blood flow of the blood vessels before and after laser irradiation is obtained, so as to evaluate the final laser treatment effect.

[0075] As shown in Figure 4 The deep learning denoising algorithm is preferably used for real-time denoising of the speckle blood flow image, which greatly improves the signal-to-noise ratio of the blood flow image. Only 5 frames of original laser speckle blood flow image can realize high-quality blood flow imaging, and the time resolution is improved. The Nd:YAG treatment laser is transmitted through the mouse spine window 108 on the intact skin side to perform laser irradiation. Preferably, the laser incident energy density is 4.8 J / cm2, the frequency is 4 Hz, and the pulse width is 1 ms. It is difficult to cause thermal damage to blood vessels. Under the same laser parameters, the blood vessels are linearly contracted and completely thermally damaged by directly irradiating the laser on the peeled skin side by using the traditional method.

[0076] Specifically, both the noise reducer and the discriminator adopt the efficient Unet architecture.

[0077] The denoiser consists of an encoder, a bridge, and a decoder. The encoder unit employs three sequentially arranged lightweight residual blocks with a stride of 1 or 2 for feature extraction, and the number of filters is doubled after downsampling. This encoder architecture is renowned for its deep learning efficiency, integrating residual blocks using depthwise separable convolutions for efficient denoising prior modeling. To ensure a large activation map, downsampling with a stride of 1 in the first layer is deferred. The extracted features are passed to the decoder via the bridge for estimating the denoised input in the feature domain. Features are transformed to the image domain by each decoder layer through a transformation unit consisting of n residual blocks followed by a 1 × 1 convolutional layer and a sigmoid activation layer. The number of filters in the transformation unit is halved n times, and the filter size of the 1 × 1 convolutional layer is set to C, i.e., the number of input channels. Unless otherwise specified, the kernel size and dilation rate of the convolutional layers remain at 3 and 1, respectively. A vascular structure enhancer is designed to preserve sparse information unaffected by noise. The vascular structure enhancer incorporates a large-kernel depthwise segregating convolution to extract low-level features, which are then captured using a set of parallel depthwise segregating convolutions to capture structural information at different scales. The structural features are fused through two 1×1 convolutional layers, and the vascular structure enhancer is appended to the last decoder layer before entering the transformation unit. The denoised data is the output of this transformation unit. This lightweight denoiser has only 1.19 million parameters, contributing to faster model inference and supporting its deployment on mobile devices.

[0078] The discriminator also employs the efficient Unet architecture, but differs in that its network consists of ordinary convolutional layers to ensure the discriminator's accuracy. In the discriminator, the output of the transformation unit is converted into a binary mask by a mapper composed of convolutional layers and sigmoid activation layers.

[0079] Figure 5 In image (a), due to the concentrated laser beam, the area affected by the laser under direct irradiation is small, resulting in a thermal effect only on the blood vessel region without significant impact on surrounding tissues. However, in... Figure 5 In the image of the laser action process in (b), the laser light is transformed into diffuse light after being scattered by the upper skin tissue. Its effective range is larger, but the beam is more divergent, reducing the therapeutic effect on blood vessels. At the same time, a large area of ​​surrounding normal tissue is affected by the laser thermal effect. This indicates that traditional methods cannot simulate the laser-induced vascular thermal effect in the presence of skin, lacking realistic guidance. Therefore, based on this method, the treatment laser can be irradiated from the intact skin side, while the other side is observed in real time using laser speckle contrast imaging technology. This provides a more realistic picture of the laser thermal effect and offers guidance for clinical treatment.

[0080] A second object of the present application is to provide a laser heat effect evaluation system, comprising:

[0081] An acquisition module is configured to acquire a sequence of original laser speckle blood flow images, analyze the original laser speckle blood flow images by using a time contrast analysis algorithm, and acquire a laser speckle contrast image;

[0082] A calculation module is configured to calculate a half-height width of a blood vessel in the laser speckle contrast image as a width of the blood vessel, and acquire a difference image by subtracting a blood flow image after laser irradiation from a blood flow image before laser irradiation;

[0083] An evaluation module is configured to acquire blood vessels with changed blood flow due to laser irradiation according to the difference image, select a region of interest, acquire a mean value of blood flow of the blood vessels before and after laser irradiation, and evaluate a final laser treatment effect.

[0084] As an example, a noise reduction module is further included:

[0085] The noise reduction module uses deep learning to reduce noise in real time for the original laser speckle blood flow image and converts the original laser speckle blood flow image into a speckle blood flow image;

[0086] The deep learning is a generative adversarial denoising method for laser speckle blood flow image denoising, a deep neural network of the generative adversarial denoising method for laser speckle blood flow image denoising is composed of a denoiser and a discriminator, and the denoising is realized by mutual confrontation between the denoiser and the discriminator;

[0087] The denoiser and the discriminator both adopt an Unet architecture; the denoiser is composed of an encoder, a bridge, and a decoder; the encoder unit adopts three lightweight residual blocks arranged in sequence, with a stride of 1 or 2, for feature extraction, and the number of filters is doubled after down-sampling operation; wherein the down-sampling with a stride of 1 of the first layer is postponed; the extracted features are transmitted to the decoder through the bridge for estimating the denoised input in the feature domain; the features are converted to the image domain by each decoder layer through a conversion unit, the conversion unit is composed of n residual blocks, followed by a 1x1 convolution layer and a Sigmoid activation layer; the number of filters in the conversion unit is halved n times, and the filter size of the 1x1 convolution layer is set to C;

[0088] A blood vessel structure enhancer is further included, the blood vessel structure enhancer includes a large-core depthwise separable convolution to extract low-level features, and then captures structure information of different scales through a group of parallel depthwise separable convolutions; the structure features are fused through two 1x1 convolution layers, and the blood vessel structure enhancer is attached to the last decoder layer before entering the conversion unit;

[0089] The discriminator network is composed of ordinary convolutional layers, and the output of the conversion unit in the discriminator is converted into a binary mask through a mapper composed of a convolutional layer and a sigmoid activation layer.

[0090] As an example, in the acquisition module, the time contrast analysis algorithm introduces a contrast signal-to-noise loss function, and the blood vessels and tissues are enhanced through a guided adversarial distortion learning network.

[0091] As an example, in the acquisition module, the laser speckle contrast image is converted to obtain a laser speckle blood flow index image by using a sequence formed by the original laser speckle blood flow image.

[0092] The system is based on the above laser heat effect evaluation method.

[0093] A third object of the embodiments of the present application is to provide an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the laser heat effect evaluation method.

[0094] A fourth object of the embodiments of the present application is to provide a computer readable storage medium, which stores a computer program, wherein the computer program is executed by a processor to implement the laser heat effect evaluation method.

[0095] A fifth object of the embodiments of the present application is to provide a computer program product, which comprises computer instructions, and the computer instructions instruct a computer to execute the laser heat effect evaluation method.

[0096] These computer program instructions can also be stored in a computer readable memory capable of guiding a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce a product comprising instruction means, which realizes the functions specified in one or more flows and / or blocks. Figure 1 one or more flows and / or blocks. Figure 1 one or more blocks.

[0097] These computer program instructions can also be loaded into a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to produce a computer implemented process, so that the instructions executed on the computer or other programmable device provide steps for realizing the functions specified in one or more flows and / or blocks. Figure 1 one or more flows and / or blocks. Figure 1 one or more blocks.

[0098] The present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment containing both software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, RAM, ROM, optical storage etc.) embodying computer readable program code.

[0099] The present application is described in reference to the flowchart and / or block diagrams of the method, apparatus (system) and computer program product according to embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagrams, as well as a combination of flows and / or blocks in the flowchart and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing apparatus to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing apparatus produce a machine that implements the flow Figure 1 one or more flows and / or blocks Figure 1 an apparatus with a function specified in one or more flows or blocks.

[0100] Obviously, the described embodiments are only some embodiments but not all embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work should belong to the scope of protection of the present application.

[0101] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application but not to limit it. Although the present application has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: the specific embodiments of the present application can be modified or replaced by the same, without departing from the spirit and scope of the present application. Any modification or equivalent replacement, which does not depart from the spirit and scope of the present application, should be covered within the protection scope of the claims of the present application.

Claims

1. A laser speckle blood flow imaging device, characterized in that, Includes a laser speckle imaging system and a therapeutic laser; the laser speckle imaging system consists of an imaging laser, a beam expander, an imaging optical system, an image acquisition system, a bandpass filter, a neutral density adjustable attenuator, and a reflector; The rat spinal window animal model was fixed in the laser speckle imaging system by a fixation device. The laser generated by the treatment laser was reflected by the first reflector and then incident perpendicularly to the intact skin side of the rat spinal window. At the same time, the blood flow on the side with the skin removed was monitored by the laser speckle imaging system. The imaging laser is a near-infrared laser source, and the laser power is adjusted by changing the angle of the neutral density adjustable attenuator. The adjusted beam is expanded by the beam expander and then illuminates the mouse spine window through the second reflector. The imaging optical system is used to image the mouse spine window, and the image is acquired by the detector and then transmitted to the data acquisition and analysis system for processing.

2. The laser speckle blood flow imaging device according to claim 1, characterized in that, It also includes installing an interference bandpass filter with the same center wavelength as the imaging laser in front of the detector to collect dynamic scattering signals; the therapeutic laser uses a therapeutic laser to irradiate the intact skin side through the mouse spinal window.

3. A laser thermal effect evaluation system, characterized in that, include: The acquisition module is used to treat subcutaneous blood vessels on the side of intact skin in the spinal window of a mouse treated with transdermal laser irradiation. On the contralateral side, laser speckle blood flow imaging is used to monitor blood flow changes in real time during the treatment process. The module acquires the original laser speckle blood flow image sequence and analyzes the original laser speckle images using a time contrast analysis algorithm to obtain laser speckle contrast images. The laser speckle contrast images are then converted to obtain laser speckle blood flow index images. The calculation module is used to calculate the half-width at half-maximum (WHM) of blood vessels in the laser speckle blood flow index image as the width of the blood vessels; and to obtain the difference image by subtracting the blood flow image after laser irradiation from the blood flow image before laser irradiation. The evaluation module is used to obtain the blood vessels that show changes in blood flow due to laser irradiation based on the difference image, then select the region of interest, obtain the average blood flow of the blood vessels before and after laser irradiation, and thus evaluate the final laser treatment effect.

4. The laser thermal effect evaluation system according to claim 3, characterized in that, After acquiring the original laser speckle blood flow image sequence, the method further includes: Deep learning is used to denoise the original laser speckle blood flow image in real time and convert it into a speckle blood flow image; The deep learning method is a laser speckle blood flow image denoising generation adversarial denoising method. The deep neural network of this laser speckle blood flow image denoising generation adversarial denoising method consists of a denoiser and a discriminator. The denoiser and the discriminator compete against each other to achieve denoising. Both the denoiser and discriminator adopt the Unet architecture. The denoiser consists of an encoder, a bridge, and a decoder. The encoder unit uses three lightweight residual blocks arranged in sequence with a stride of 1 or 2 for feature extraction, and the number of filters is doubled after downsampling. The downsampling with a stride of 1 in the first layer is deferred. The extracted features are passed to the decoder through the bridge to estimate the denoised input in the feature domain. The features are transformed to the image domain by each decoder layer through a transformation unit. The transformation unit consists of n residual blocks, followed by a 1×1 convolutional layer and a sigmoid activation layer. The number of filters in the transformation unit is halved n times, and the filter size of the 1×1 convolutional layer is set to C. It also includes a vascular structure enhancer, which contains a large kernel depthwise segregating convolution to extract low-level features, and then captures structural information at different scales through a set of parallel depthwise segregating convolutions; the structural features are fused through two 1×1 convolutional layers, and the vascular structure enhancer is attached to the last decoder layer before entering the conversion unit. The discriminator network consists of ordinary convolutional layers. The output of the transformation unit in the discriminator is converted into a binary mask by a mapper, which consists of convolutional layers and sigmoid activation layers.

5. The laser thermal effect evaluation system according to claim 3, characterized in that, The temporal contrast analysis algorithm introduces a contrast signal-to-noise ratio loss function and enhances the contrast between blood vessels and tissues by guiding an adversarial distortion learning network.

6. The laser thermal effect evaluation system according to claim 3, characterized in that, The process of converting the laser speckle contrast image to obtain the laser speckle blood flow index image is a sequence formed using the original laser speckle blood flow image.

7. An electronic device, characterized in that, The system includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the evaluation method of the laser thermal effect evaluation system according to any one of claims 4-6.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the evaluation method of the laser thermal effect evaluation system according to any one of claims 4-6.

9. A computer program product, the computer program product comprising computer instructions, characterized in that, The computer instructions instruct the computer to execute the evaluation method of the laser thermal effect evaluation system according to any one of claims 4-6.

Images